Description

AI Team – AI Sound-Based Violence Detection System 


Introduction 

This project will develop an AI-powered sound analysis system designed to detect audio patterns that may indicate violent or dangerous situations in public spaces.
The system will use microphones connected to the smart campus / municipal network to continuously monitor environmental sounds, analyze them in real time, and identify audio events such as shouting, screams, gunshots, glass breaking, or aggressive crowd noise.
Detected threats will trigger intelligent alerts sent to a control center dashboard, enabling faster response and improved urban safety.
Background 
Urban security systems rely heavily on visual surveillance, which can be limited by lighting conditions, camera angles, or obstructions.
Audio-based threat detection provides an additional and complementary layer of situational awareness.
Recent advances in AI and deep learning for audio signal processing enable accurate classification of complex sound events. By integrating audio sensors into the smart campus infrastructure, the system can detect potential violent incidents even when cameras fail to capture them clearly.

Project Scope 

• Stream real-time audio data to the AI processing server.
• Preprocess audio signals (noise reduction, segmentation, feature extraction).
• Deploy AI models to classify sound events and detect potentially violent audio patterns.
• Send automated alerts with timestamps, sound type, and confidence level to the control center dashboard.





Student Requirements 

• Proficiency in Python.
• Basic knowledge of signal processing and audio concepts.
• Familiarity with machine learning / deep learning fundamentals.
• Experience with data preprocessing and model training.
• Independent learning and high motivation.
Development Tools 
• Python
• Audio Processing Libraries: Librosa, NumPy, SciPy
• Machine Learning / AI: PyTorch, TensorFlow, Scikit-learn
• Messaging & Communication: MQTT / REST APIs
• Databases: MongoDB (preferred)
• Additional Tools: GitHub, VS Code, Postman
Deliverables 
• Trained AI model for violent and abnormal sound detection.
• Specification document. 
• YouTube video 
• Poster 
• Presentation 
• GitHub link to the code 
Name: Shimon Turchak
Email: shimonturchak2@gmail.com
AI Sound-Based Violence Detection

AI and Machine Learning

Description

Emphasis in project execution

The project is has cooperation with the industry and combines meeting deadlines while being creative and focused on the task
